Vol. 29, No. 01 (2023), pp. 36β44.
NUMERICAL INVARIANTS OF COPRIME GRAPH OF A GENERALIZED QUATERNION GROUP
Nurhabibah1,a, I Gede Adhitya Wisnu Wardhana1,b, and Ni Wayan Switrayni1,c
1Department of Mathematics, Faculty of Mathematics and Natural Science, University of Mataram. Email:
bcorresponding author: [email protected]
Abstract.The coprime graph of a finite group was defined by Ma, denoted by ΞG, is a graph with vertices that are all elements of groupGand two distinct verticesx andyare adjacent if and only if (|x|,|y|) = 1. In this study, we discuss the numerical invariants of a generalized quaternion group. The numerical invariant is a property of a graph in numerical value and that value is always the same on an isomorphic graph. This research is fundamental research and analysis based on patterns in some examples. Some results of this research are the independence number of ΞQ4n
is 4nβ1 or 3nand its complement metric dimension is 4nβ2 for eachnβ₯2.
Key words: coprime graph, generalized quaternion group, numerical invariants
1. INTRODUCTION
The graph representation of an algebraic structure becomes a hot topic in recent years. For a finite groupG, we can representGin some simple graphs such as the coprime graph, the non-coprime graph, the power graph, the intersection graph, the commuting graph, and others. See [1] [2] [4] [11] [12] for details. In 2014, Ma et al [1] introduced the coprime graph of a finite group G, where the vertex set of the graph is Gand two distinct vertices xandy are adjacent if and only if the order of xand the order ofy are relative primes [1].
In 2021, Nurhabibah et al. [4] give some results on the shape of the coprime graph of a generalized quaternion group as a bipartite graph, tripartite graph, or multipartite graph. So, in this article, we would like to discuss numerical invariants
2020 Mathematics Subject Classification: 05C25 Received: 04-07-2022, accepted: 22-02-2023.
36
of the graph as degree, radius, diameter, domination number, independence num- ber, girth, metric dimension, and complement metric dimension. The numerical invariant is a property of a graph in numerical value and that value is always the same on an isomorphic graph.
2. DEFINITION AND SOME PROPERTIES
A generalized quaternion group is a special group with a definition as follows.
Definition 2.1. [5] A Generalized quaternion group (Q4n)with nβ₯2 is a group with a presentation
< a, b|a2n =e, an=b2, bβ1ab=aβ1>
in this group, akb=baβk and the order ofakbis 4.
For example,D6={e, a, a2, b, ab, a2b}is a dihedral group with order six. In this study, we give a functionf that defines the adjacency of two vertices as follows.
Definition 2.2. GivenH = (V, E)is a graph withV 6=β . Define f as follows.
f :V ΓV β {0,1}
with
f((vi, vj))
1, if(vi, vj)βE
0, else
One graph that is associated with a finite group is the coprime graph. The definition of this graph is given in Definition 2.3.
Definition 2.3. [1] Given G a finite group, the coprime graph of G, denoted by ΞG is the graph with V(ΞG) =Gand two distinct verticesx andy are adjacent if and only if(|x|,|y|) = 1.
And now, we define some numerical invariants of the graph that we analyze in this study.
Definition 2.4. Let G be a graph. The degree of vertex v in G is denoted as deg(v) =|A| whereA={viβV|f((v, vi)) = 1}.
Definition 2.5. [6] Let G be a graph and c(v) = max{d(v, x)|x β V(G)} is the eccentricity of G, rad(G) =min{c(v)|v βV(G)} and diam(G) = maks{c(v)|v β V(G)}.
Definition 2.6. [7]Let G= (V, E) be a graph. The girth of G, denoted byg(G) is defined as:
g(G) =
min{ |CG||CG cycle ofG}, Gcontains cycle
0, else
Where|CG| is the length of the cycleCG.
Definition 2.7. [8] Let G be a graph, the domination number of G, denoted by Ξ³(G) is the minimum cardinality of the domination set ofG. The domination set is a subset D of V(G) such that every vertex not inD is adjacent to at least one member of D.
Definition 2.8. [9] Let Gbe a graph, the independence number of G, denoted by Ξ²(G)is the maximum cardinality of the independence set ofG. The independence set is a subset I of V(G) such that no two vertices in the subset represent an edge of V(G).
Definition 2.9. [10] Let G be a graph, the metric dimension of G, denoted by dim(G)is the minimum cardinality of the resolving set of G. The resolving set is a subset W for a graph G if, for every two distinct verticesuand v of G, there is an element win W that resolves uandv.
Definition 2.10. [11] Let G be a graph, complement metric dimension ofG, de- noted by dim(G) is the maximum cardinality of complement resolving a set ofG.
Definition 2.11. LetΞGbe the coprime graph of a finite groupGandxβV(ΞG).
We can defineAxas Ax={xiβV(ΞG)|f((x, xi)) = 1}.
We will discuss numerical invariants of the generalized quaternion group that is represented in the coprime graph. Given the previous result of the coprime graph of a generalized quaternion group.
Theorem 2.12. [4]Let Q4n be a generalized quaternion group. Ifn= 2k then the coprime graph ofQ4n is a complete bipartite graph.
Theorem 2.13. [4]LetQ4n be a generalized quaternion group. Ifnis prime, then the coprime graph of Q4n is tripartite graph.
Theorem 2.14. [4]LetQ4n be a generalized quaternion group. Ifn=pk11pk22...pkmm, pi6= 2, pi 6=pj for each i6=j, pi a prime number then the coprime graph ofQ4n is m+ 2partite graph.
Theorem 2.15. [4]LetQ4n be a generalized quaternion group. Ifn=pk11pk22...pkmm, p1= 2,pi are distinct prime numbers then the coprime graph ofQ4nism+1partite graph.
We also give the following theorem that we need to prove some numerical invariants of the coprime graph of a finite group.
Theorem 2.16. Let (G,β) be a finite group, |G| = n =pk11pk22...pkmm and ΞG be the coprime graph of G. For some x, y β V(ΞG) with x 6= y, 1 β€ li, ti β€ ki,
|x|=Qj
i=1plii and|y|=Qj
i=1ptii if and only if Ax=Ay.
Proof. Take any x1 β Ax, it means (|x1|,|x|) = 1. Note that |x| = Qj i=1plii and (|x1|,|x|) = 1, so 6 βpi with 1 β€i β€j such that pi||x1|. Since |y|=Qj
i=1ptii and 6 βpi with 1 β€i β€ j which resulted in pi||x1| then (|y|,|x1|) = 1 and hence f((x1, y)) = 1. Thus,x1βAy, which means AxβAy. By a similar approach, we can prove thatAy βAx, henceAx=Ay
LetGbe a finite group with|G|=n=pk11pk22...pkmm and ΞG be the coprime graph of G. Define A = {1,2, ..., m} and B, C β A with B 6= C. If x, y β G,
|x|=Q
sβBplss and|y| =Q
rβCplrr , we need to prove Ax 6=Ay. Choosez β Ax with|z|=pa whereaβC, buta /βB. Consequently, (|y|,|z|) =pa6= 1. It means z /βAy. So, it is complete to prove Ax6=Ay.
3. NUMERICAL INVARIANTS OF THE COPRIME GRAPH OF Q4n
In this section, we analyze the numerical invariants of the coprime graph of Q4n. The first numerical invariants obtained from this study are the degree of each vertex as stated in Theorem 3.1 and Theorem 3.2.
Theorem 3.1. Let ΞQ4n be the coprime graph of Q4n. Ifn= 2k withkβ₯1 then deg(e) = 4nβ1 anddeg(v) = 1for each vβQ4n\{e}.
Proof. Let ΞQ4n be the coprime graph ofQ4n. Taken= 2k. By Theorem 2.12, ΞQ4nis a complete bipartite graph with partitionsV1={e}andV2=Q4n\{e}, see the proof in [4] for detail. Thus,f((e, v)) = 1, for eachvβQ4n\{e}. By definition, obtained deg(e) =|V2|= 4nβ1 anddeg(v) =|V1|= 1.
And fornis an odd prime number we have
Theorem 3.2. LetΞQ4nbe the coprime graph ofQ4n. Ifnis an odd prime number, the degree of each vertex is 1, n,2n+ 2, or4nβ1.
Proof. Let ΞQ4n be the coprime graph ofQ4n. Takenfor any odd prime number p. Let S1 = {e}, S2 = {ap, b, ab, ..., a2pβ1b}, S3 = {a2, a4, ..., a2pβ2}, and S4 = {a, a3, ..., apβ2, ap+2, ..., a2pβ1}. Note that |e| = 1, |x| = 2 or |x| = 4 for each xβS2,|y|=pfor eachyβS3, and|z|= 2pfor eachzβS4. Thus,
β’ f((e, v1)) = 1 for eachv1βA1 withA1=Q4n\{e}, sodeg(e) =|A1|= 4nβ1.
β’ ForxβS2, we getf((x, v2)) = 1 if and only if v2βA2 withA2=S1βͺS3. So, deg(x) =|A2|=|S1|+|S3|=n.
β’ Fory βS3, we get f((y, v3)) = 1 if and only if vβA3 withA3=S1βͺS2. So, deg(y) =|A3|=|S1|+|S2|= 1 + 2n+ 1 = 2n+ 2.
β’ For z β S4, we get f((z, v4)) = 1 if and only if v4 β A4 with A4 = S1. So, deg(z) =|S1|= 1.
The next numerical invariants are radius and diameter.
Theorem 3.3. IfΞQ4n is the coprime graph of a generalized quaternion group with nβ₯2, thenrad(ΞQ4n) = 1anddiam(ΞQ4n) = 2.
Proof. Let ΞQ4n be the coprime graph ofQ4n. Take anynβNandnβ₯2. From [4] we know that 1β€ d(u, v)β€2 for each u, v β V(ΞQ4n). It meansc(v) = 1 or c(v) = 2 and by definition we get rad(ΞQ4n) = 1 anddiam(ΞQ4n) = 2.
The next result obtained from this study is the girth of ΞQ4nas stated in the two following theorem.
Theorem 3.4. IfΞQ4n is the coprime graph of a generalized quaternion group with n= 2k theng(ΞQ4n) = 0.
Proof. According to Theorem 2.12, if n= 2k then ΞQ4n is a complete bipartite graph. It means, there is no cycle in ΞQ4n. By definition, we getg(ΞQ4n) = 0.
And for other cases we have
Theorem 3.5. IfΞQ4n is the coprime graph of a generalized quaternion group with n6= 2k andn >2 theng(ΞQ4n) = 3.
Proof. Let ΞQ4n is the coprime graph of a generalized quaternion group Q4n. Take n β Nwith n > 2 and n 6= 2k. There arev1, v2 βV (ΞQ4n) with|v1|= p, p an odd prime number and |v2| = 2. So, we get a cycle with length 3 which is eβv1βv2βv1. It is complete to proveg(ΞQ4n) = 3.
The domination number is one of the numerical invariants of the graph. The domination number of ΞQ4n is always the same for all n β₯ 2. This property is stated in Theorem 3.6.
Theorem 3.6. If ΞQ4n be the coprime graph of a generalized quaternion then Ξ³(ΞQ4n) = 1for each nβ₯2.
Proof. LetD={e} andD be the domination set because eis adjacent to other vertices in ΞQ4n. So,Ξ³(ΞQ4n) =|D|= 1.
The next numerical invariant is the independence number of the graph. This property is stated in four following Theorem.
Theorem 3.7. Let ΞQ4n be the coprime graph of a generalized quaternion group.
If n= 2k thenΞ²(ΞQ4n) = 4nβ1.
Proof. Let ΞQ4n be the coprime graph of a generalized quaternion group with n = 2k. By Theorem 2.12, ΞQ4n is a complete bipartite graph that is a star graph. It means there are two partitions and these partitions are independent sets, V1 = {e} and V2 = Q4n\{e}. Now, suppose that there is independence set I with |Q4n| β₯ |I| > |V2|. So, I must equal to Q4n and f((e, v)) = 1 for any v β Q4n\{e}. It is a contradiction and it means V2 is an independence set with maximum cardinality. Finally, we getΞ²(Q4n) =|V2|= 4nβ1.
And for casenis an odd prime we have
Theorem 3.8. Let Q4n be a generalized quaternion group, n=p,pan odd prime number, thenΞ²(ΞQ4n) = 3n.
Proof. Let ΞQ4nbe the coprime graph ofQ4n. Wherenis an arbitrary odd prime number. Define I = {aj|j = 1,3, ...,2nβ1} βͺ {aib|i = 0,1,2, ...,2nβ1}. Note that |v|= 2q, qβN,βvβI. Thus, βvi, vj βI it is valid that (|vi|,|vj|)6= 1 which resulted inf((vi, vj)) = 0. So,I is the independence set of ΞQ4n. And we need to prove that there is no independence set of ΞQ4n with condition|I0|>|I|= 3n. If there isI0, we can findvβI0, butv /βI, it means|v|= 1 or|v|=p. Consequently, there isuβIβ©I0 with condition (|u|,|v|) = 1 which means f((u, v)) = 1. It is a contradiction withI0 independence set of ΞQ4n. So,Ξ²(ΞQ4n) =|I|= 3n.
And for a more general case, whennis an odd
Theorem 3.9. LetΞQ4nbe the coprime graph ofQ4n. Ifn=pk11pk22...pkmm,pi6=pj, i6=j,pi an odd prime number, thenΞ²(ΞQ4n) = 3n.
Proof. Let ΞQ4nbe the coprime graph ofQ4n. Taken=pk11pk22...pkmm,pi6= 2,pi6=
pj,i6=j,pian odd prime number. DefineI=I1βͺI2withI1={a2lβ1|l= 1,2, ..., n}
and I2 = {aib|i = 0,1,2, ...,2nβ1}. Take a2lβ1 β I1, we get |a2lβ1| = 2z1 with z1 βN andβaibβI2 we get |aib|= 4. Thus,βu, v βI we get (|u|,|v|) = 2z2 6= 1 which shows that f((u, v)) = 0. So,I is an independence set. Suppose that I0 is an independence set with|I0|>|I|= 3n. We can findv1βI0 andv1β/I, it means
|v1| 6= 2z3, βz3 βZ. Because of |I0| >3n then |Iβ©I0| β₯ 2n+ 1. Consequently,
βv2 β Iβ©I0 with v2 = aib with i β {0,1, ...,2nβ1}, it means |v2| = 4. Thus, (|v1|,|v2|) = 1 orf((v1, v2)) = 1. It is a contradiction withI0 is independence set.
So,Ξ²(ΞQ4n) =|I|= 3n.
And for a more general result
Theorem 3.10. IfΞQ4n be the coprime graph ofQ4n. Ifn=pk11pk22...pkmm,p1= 2, pi6=pj,i6=j thenΞ²(ΞQ4n) = 4nβ2nk1.
Proof. Let ΞQ4n be the coprime graph of Q4n. Taken= pk11pk22...pkmm, p1 = 2, pi6=pj,i6=j. DefineI=I1βͺI2βͺI3withI1={a2lβ1|l= 1,2, ..., n},I2={a2l|l=
1,2, ..., nβ1}\{a2k1 +1q|q = 1,2, ...,2nk1 β1}, and I3 ={aib|i= 0,1,2, ...,2nβ1}.
For any vβI we get|v|= 2z1, z1βN. Consequently,βu, v βI we get (|u|,|v|) = 2z26= 1. ThusI is an independence set. And now, we need to prove that there is noI0 as the independence set of ΞQ4nwith condition|I0|>|I|. Suppose that there is I0 with this condition, thenβv1βI0 andv1 β/ I, It means|v1| 6= 2z3, βz3 βZ.
Because of|I0|>|I|then|Iβ©I0| β₯4nβ2kn1β1+ 1>3n. Consequently,βv2βIβ©I0 wherev2=aiband|v2|= 4. Those, we get (|v1|,|v2|) = 1 orf((v1, v2)) = 1. It is a contradiction. So,Ξ²(ΞQ4n) =|I|=|I1|+|I2|+|I3|=n+ (nβ1)β 2nk1 β1
+ 2n= 4nβ2nk1.
Two of the numerical invariants of the graph that are defined based on the metric concept are the metric dimension and the complement metric dimension.
In this study, we find the metric dimension and complement metric dimension of ΞQ4n.
Theorem 3.11. LetΞQ4nbe the coprime graph ofQ4n. Ifn= 2k thendim(ΞQ4n) = 4nβ2.
Proof. Let ΞQ4n be the coprime graph of Q4n. If n = 2k, k β N, then ΞQ4n is star graph. Choose W = Q4n\{e, a} = {v1, v2, ..., v4nβ2},βv β W obtained r(vi|W) = (d1, d2, ..., d4nβ2) with di = 0 and dj = 2 for each i 6= j. So, each v β W has a distinct representation of W. And then r(e|W) = (1,1, ...,1) and r(a|W) = (2,2, ...,2). Thus, each vertex of ΞQ4n has distinct representation toW. In other words,W is a resolving set of ΞQ4n.
Suppose that there isW0 as a resolving set with condition |W0| < 4nβ2.
There are two distinct vertices v1, v2 β Q4n\{e} and v1, v2 β/ W0 which resulted in r(v1|W0) = r(v2|W0). Thus, W0 is not resolving set and it is complete to dim(ΞQ4n) =|W|= 4nβ2.
And fornis an odd prime number
Theorem 3.12. Let ΞQ4n be the coprime graph of Q4n. Ifn=p,pan odd prime number, thendim(ΞQ4n) = 4nβ4.
Proof. Let ΞQ4n be the coprime graph ofQ4n. Letnbe any odd prime number.
Choose W =Q4n\{e, a, a2, b}. Note that eachv1, v2 βW with v1 6=v2 obtained r(v1|W)6=r(v2|W). Because of|e| 6=|a| 6=|a2| 6=|b|and the set of prime numbers that divide the order ofe, a, a2, andb are not equal, so by Theorem 2.16 obtained Ae 6=Aa 6=Aa2 6=Ab. So, r(e|W)6=r(a|W)6=r(a2|W)6=r(b|W). Thus W is a resolving set of ΞQ4n.
Suppose that there isW0 with condition|W0|<|W|= 4nβ4 as a resolving set of ΞQ4n. We can find two distinct vertices u, v /β W0 with |u| = |v|, means Au = Av and r(u|W0) = r(v|W0). Thus, W is a resolving set with minimum cardinality. In other words,|Wβ| β₯4nβ4 for eachWβ resolving set of ΞQ4n. So, dim(ΞQ4n) =|W|= 4nβ4.
And fornis an odd number
Theorem 3.13. Let ΞQ4n be the coprime graph ofQ4n. Ifn=pk11pk22...pkmm,pi an odd prime number,pi6=pj fori6=j then dim(ΞQ4n) = 4nβ2m+1.
Proof. Let ΞQ4n be the coprime graph of Q4n. Take n = pk11pk22...pkmm, pi 6=
pj for i 6= j. Define W = Q4n\S with S = {e} βͺS1βͺ...βͺSm+1 and St = {vt1, vt2, ..., vtCm+1
t } for each 1 β€tβ€m+ 1 with |vts| has t distinct prime factor and there are no two distinct elements inStwithtthe same prime factor. Thus,|St| is the total of ways to choosetprime numbers fromm+ 1 distinct prime numbers or|St|=Ctm+1 where Ckm=(mβk)!k!m! . Consequently,
|S|= 1 +C1m+1+C2m+1+...+Cm+1m+1= 2m+1
And now, we need to proveW =Q4n\S is the resolving set. Eachv1, v2βW with v1 6= v2, obtained r(vi|W) = (d1, d2, ..., d|W|) with di = 0 and dj 6= 0 for eachi6=j, thus r(v1|W)6=r(v2|W). By Theorem 2.16, each two distinct vertices u1, u2 β S, we get Au1 6= Au2 which resulted in r(u1|W) 6= r(u2|W). Thus, r(u|W)6=r(v|W) for eachu, vβW with u6=v. In other words,W is a resolving set of ΞQ4n.
Suppose that W0 = Q4n\S0 with |S0| > |S| is resolving set of ΞQ4n. We can find v5, v6 βS0 withv56=v6, but |v5| and |v6| have t the same prime factor.
By Theorem 2.16, obtained Av5 =Av6 and r(v5|W0) = r(v6|W0). So, W0 is not resolving set andW is resolving set ΞQ4n with minimum cardinality. It is done to provedim(ΞQ4n) =|W|= 4nβ2m+1.
And for a more general case, by Theorem 3.13 we have the following result.
Corollary 3.14. Let ΞQ4n be the coprime graph of Q4n. If n = pk11pk22...pkmm, p1= 2,pi6=pj fori6=j,pi prime number, thendim(ΞQ4n) = 4nβ2m.
Proof. Let ΞQ4n be the coprime graph of Q4n. Taken= pk11pk22...pkmm, p1 = 2, pi 6= pj for i 6= j, pi prime number. Note that p1 = 2, so the number of odd prime number pj is mβ1. Consequently, Theorem 3.13 obtained dim(ΞQ4n) = 4nβ2(mβ1)+1= 4nβ2m.
The last theorem explains the complement metric dimension of ΞQ4n.
Theorem 3.15. IfΞQ4nbe the coprime graph ofQ4nwithnβ₯2thendim(ΞQ4n) = 4nβ2.
Proof. Let ΞQ4n be the coprime graph of Q4n. Choose any natural number n with n β₯ 2. Choose S = Q4n\{b, ab}, obtained r(b|S) = r(ab|S). So, S is a complement to resolving a set of Q4n. By definition, it is impossible to find a complement resolving set with cardinality greater than |V(G)| β2. Thus, S is a complement resolving set with maximum cardinality. It is done to prove that dim(ΞQ4n) =|S|= 4nβ2.
4. CONCLUSION
Some numerical invariants of the coprime graph of a generalized quaternion that were obtained from this study are the degree of each vertex is 1, n,2n+ 2, or 4nβ1, its radius is 1, its diameter is 2, its girth is 0 or 3, its independence number is 4nβ1, 3n, or 4nβ2nk1, its domination number is 1, its metric dimension is 4nβ2,4nβ4,4nβ2m, or 4nβ2m+1, and its complement metric dimension is 4nβ2.
REFERENCES
[1] X. Ma, H. Wei, and L.Yang, The Coprime Graph of a Group, International Journal of Group Theory,β 13-23 (2014).
[2] A.G. Syarifudin, Nurhabibah, D.P. Malik, and I.G.A.W. Wardhana, Some Characterizations of the Coprime Graph of Dihedral GroupD2n,Journal of Physics: Conference Series, Vol.
1722, No. 1, p. 012051 (2021).
[3] R. Juliana, Masriani, I.G.A.W. Wardahana, N.W. Switrayni, and Irwansyah, Coprime Graph of Integer Modulo n Group and its Subgroup, Journal of Fundamental Mathematics and Applications, 15-18 (2020).
[4] Nurhabibah, A.G. Syarifudin, and I.G.A.W.Wardhana, Some Results of The Coprime Graph of a Generalized Quaternion GroupQ4n,Indonesian Journal of Pure and Applied Mathe- matics, 29-33 (2021).
[5] B. Ahmadi and H. Doostie, On the Periods of 2-Steps General Fibonacci Sequences in a Generalized Quaternion Groups,Discrete Dynamics and Nature Society, 1-8 (2012).
[6] S. Zahidah, D.M. Mahanani, K.L. Oktaviana. Connectivity Indices Of Coprime Graph Of- generalized Quaternion Group. InJournal of The Indonesian Mathematical SocietyVol. 27, No. 3, 285296 (2021).
[7] R. Diestel,Graph Theory. (Springer: 2000).
[8] N. Murugesan and D.S. Nair, The Domination and Independence of Some Cubic Bipartite Graphs,Int. J. Contemp. Matth.Sciences, 611-618 (2011).
[9] G. Chartrand, L. Eroh, M.A. Johnson, and O.R. Oellerman, Resolvability in Graphs and the Metric Dimension of a Graph,Discrete Applied Mathematics, 99-113 (2000).
[10] L. Susilowati, A.R. Slamin, and Rosfiana, The Complement Metric Dimension of Graphs and its Operations,.Int. J. Civ. Eng. Technol. (IJCIET), 2386-2396 (2019).
[11] E.Y. Asmarani, A.G. Syarifudin, I.G.A.W. Wardhana, and N.W. Switrayni. The Power Graph of a Dihedral Group.Eigen Mathematics Journal, 80-85 (2021).
[12] W.U. Misuki, I.G.A.W. Wardhana, N.W. Switrayni, and Irwansyah. Some results of non- coprime graph of the dihedral group D 2 n for na prime power. InAIP Conference Proceedings Vol. 2329, No. 1, p. 020005 (2021).
[13] Nurhabibah, D.P. Malik, H. Syafitri, and I.G.A.W. Wardhana. Some results of the non- coprime graph of a generalized quaternion group for some n. InAIP Conference Proceedings Vol. 2641, No. 1, p. 020001 (2022).